China
Africa
Explore chapters and articles related to this topic
Test and Verification
Published in Miroslav Popovic, Communication Protocol Engineering, 2018
Generally, leader election is a fundamental problem in distributed systems, because many hard-distributed problems are easy to solve once a central coordinator is available. An attractive approach to solve the leader election is by using self-stabilizing algorithms, which do not require initialization in order to operate correctly, and which can recover from transient faults that may destroy the system state information. Also, among many models, a network of finite-state anonymous agents is a rather interesting one, because it models many distributed systems of identical, simple computational nodes, such as wireless sensor networks, etc. It is well-known that the self-stabilizing leader election is impossible without a failure detector, which is a kind of oracle that provides some information to the system that it is unable to compute on its own.
Clustering Techniques
Published in Amine Dahane, Nasr-Eddine Berrached, Mobile, Wireless and Sensor Networks, 2019
Amine Dahane, Nasr-Eddine Berrached
If two nodes start the leader election process at the same time and the two nodes can hear each other, one stopped and the other continues the election process. If the two nodes cannot hear each other, each node selects a leader in its environment and at the end of the process, only one as leader is chosen (step 6, phase1).
A synod based deterministic and indulgent leader election protocol for asynchronous large groups
Published in International Journal of Parallel, Emergent and Distributed Systems, 2022
Sathyanarayanan Srinivasan, Ramesh kandukoori
Initially, the Acceptor set is predetermined and propagated to all nodes. Each node ranks the Acceptor set based on the acceptor's priority and thus all nodes have the acceptors ranked in the same order. When a node initiates leader election, it sends the Initiate message proposing itself to the acceptor with the highest rank. An acceptor on receiving a first Initiate message starts election and replies with an acknowledge message to the initiator. If an acceptor receives Initiate messages after election has started, it simply responds with an acknowledge message to the initiator. If the initiator does not receive acknowledge message from the acceptor within some time, then the node sends the Initiate message to the acceptor with the second highest rank and so on.
An election algorithm to ensure the high availability of leader in large mobile ad hoc networks
Published in International Journal of Parallel, Emergent and Distributed Systems, 2018
Shantanu Sharma, Awadhesh Kumar Singh
The Leader election (LE) is a basic challenge of distributed systems and finds a distinguished node as a leader (or coordinator) of the network. A LE algorithm may be triggered by some nodes in the network, and after its successful termination, the network has exactly one node as leader, and others are normal. More details about LE problem may be found in Chapter 11 of [16], Chapter 3 of [3], and Chapter 13 of [15]. A LE algorithm should satisfy the following properties:Safety. Eventually, all the nodes agree on a single node as a leader of the network (and no two nodes elect two different nodes as the leader). (This property is also known as an eventual leader [2,7,8].)Liveness. Every node elects a node as a leader of the network within a finite amount of time.Several reasons, e.g. dynamic nature of nodes, variation in node’s capabilities, an unsecured communication channel, and the network scalability [13], make the election process challenging in dynamic networks as compared to conventional (static) networks. Hereafter, throughout the article, the words ‘mobile ad hoc network,’ ‘system,’ and ‘network’ have been used interchangeably.
Crowd tracking and monitoring middleware via Map-Reduce
Published in International Journal of Parallel, Emergent and Distributed Systems, 2022
Alexandros Gazis, Eleftheria Katsiri
In future steps of this work, we need to focus on implementing a leader election algorithm. Specifically, a distributed algorithm must be developed to help select a process among a set of processes that will act as the leader. This leader will be responsible for the coordination of the next steps among the processes. Leader election is particularly important in distributed systems, such as sensor networks where there is no central control and each node does not know their role in the network. This way, a fair allocation of roles and tasks can be achieved.